1조6000억원 들어간 정부 'AI 학습용 데이터', 상당수는 저품질·무용지물
2026-08-12
Chosun.com
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
A government audit in South Korea found that public sector AI training datasets were often low-quality, duplicated, and lacked proper quality control, risking future AI system malfunctions. Key issues included missing essential data, lack of third-party verification, and inefficient management across multiple agencies.[AI generated]
Why's our monitor labelling this an incident or hazard?
The event involves AI systems indirectly because the data in question is intended for AI training. The poor quality and mismanagement of this data could plausibly lead to AI systems malfunctioning or producing erroneous outputs, which fits the definition of an AI Hazard. There is no description of actual harm having occurred yet, only the potential for harm due to low-quality data. The audit and recommendations are responses to this hazard but do not themselves constitute complementary information about a past incident. Hence, the classification as AI Hazard is appropriate.[AI generated]